Sociodemographic and Clinical Characteristics Associated with Worst Pain Intensity among Cancer Patients

Abstract

Objective:

Cancer patients have pain due to their cancer, the cancer treatment and other causes, and the pain intensity varies considerably between individuals. Further research is needed to understand the factors associated with worst pain intensity. Our study aim was to determine the association between worst pain intensity and sociodemographics and cancer-specific factors among cancer patients.

Methods:

A total of 1,280 cancer patients recruited from multiple cancer centers over 25 years in the United States were asked to complete a questionnaire that collected respondents’ demographic, chronic pain, and cancer-specific information. Worst, least, and current pain intensities were captured using a modified McGill Pain Questionnaire (MPQ, pain intensity measured on 0–10 scale). A generalized linear regression analysis was utilized to assess the associations between significant bivariate predictors and worst pain intensity scores.

Results:

Our study sample was non-Hispanic White (64.5%), non-Hispanic Black (28.3%), and Hispanic (7.2%). On average, participants were 59.4 (SD=14.4) years old. The average worst pain intensity score was 6.6 (SD=2.50). After controlling for selected covariates, being Hispanic (β=0.6859), previous toothache pain (β=0.0960), headache pain (β=0.0549), and stomachache pain (β=0.0577) were positively associated with worse cancer pain. Notably, year of enrollment was not statistically associated with pain.

Conclusion:

Findings identified being Hispanic and having previous severe toothache, stomachache, and headache pain as significant predictors of worst pain intensity among cancer patients. After controlling for selected covariates, we did not note statistical differences in worst pain over a 25-year period. Therefore, studies focused on improving the management of pain among patients with cancer should target interventions for those with Hispanic heritage and those with past history of severe common pain.

Introduction

Pain is one of the leading causes of disability among cancer patients with 48% of early-stage patients and 75% of late-stage patients suffering from chronic pain.^1,2 Cancer patients are more likely to endorse having pain than the general population due to several factors including tumor progression, co-morbid conditions, and side effects of their treatments.^3,4 However, contributors to an individual’s cancer pain are multifactorial and often challenging to measure. Thus, studies assessing cancer pain are generally limited to measuring the individual effects of sociodemographic and clinical characteristics within various populations.^5–8 Understanding correlates of cancer pain while controlling for sociodemographic and clinical characteristics is a crucial step for improving pain throughout the cancer control continuum.

Previous research identified racial/ethnic disparities regarding oncological outcomes including pain severity.^9–12 Non-Hispanic White individuals are less likely to report having chronic pain compared to Hispanic and non-Hispanic Black individuals.¹⁰ Notably, Cleeland et al, noted older, minority women, were at the greatest risk for poor cancer pain control compared to individuals without those factors.¹³ Studies targeting gender differences indicate women are more likely to endorse severe pain compared to men,^14,15 whereas others indicate greater pain severity scores in men compared to women.^16,17 Individuals with lower education levels are more likely to endorse having chronic pain compared to individuals with higher education.¹⁸ Similarly, multiple studies reported greater proportions of pain among non-Hispanic Black and Hispanic participants compared to white participants.^19–21 Further, one study noted Asian and non-Hispanic Black patients with esophageal and gastric cancer were more likely to report abdominal pain than white participants.²¹ Additionally, Anderson et al., did not identify statistical differences between non-Hispanic Black participants and Hispanic participants, however this study did not include non-Hispanic White participants.²² What remains unknown is whether the reported disparities in cancer pain intensity are driven by clinical or sociodemographic characteristics alone.

To date, few studies have examined associations between pain intensity and clinical characteristics of cancer including the location and tumor stage.^23,24 For example, Schlaeger et al., noted statistical differences between lung or prostate pain quality descriptors.²⁵ However, little is understood regarding differences in pain intensity across additional cancer locations including colorectal and gastrointestinal sites. One recent longitudinal analysis concluded, late-stage cancer patients reported greater pain severity compared to early-stage cancer patients.²⁶ Further, Berry et al., identified positive correlations between pain intensity and common pains, defined as stomachache, toothache, and headaches pain, among a sample of lung cancer patients.²⁷ While these findings demonstrate associations between common.^19,20*

As clinicians seek to reduce the severity of pain among patients, further research is needed to better understand the associations between common pain experiences and worst pain intensity among persons living with cancer. We are especially interested in worst pain intensity to account for recall bias associated with the peak-end bias observed in pain settings.²⁸ The peak-end bias occurs when study participants rate their “average” pain based on discomfort experienced at the peak rather than the average. Therefore, using worst pain intensity ratings may better represent pain ratings than average pain.²⁹

Therefore, the aims of this study were to (1) identify racial/ethnic differences in worst cancer pain intensity and (2) explore the relationships among specific types of cancer, common prior pain experiences including worst toothache, headache, and stomachache and worst pain intensity.

Methods

Study Sample

Participants were recruited via outpatient oncology centers and palliative care programs in Seattle, WA and Chicago, IL. Participants were included if they had a cancer diagnosis, reported pain during the past week, and spoke/read English. Participants were excluded if they had surgery less than two months prior to the study visit or were mentally or physically unable to complete study procedures.

Participants were queried about several cancer-related domains including demographics, pain, and cancer-specific domains using paper or computer versions of the survey. Researchers conducted medical record abstraction. Surveys were completed in multiple waves between the years of 1994 to 2018. Detailed methods for study enrollment were previously described.³⁰ The primary studies were approved by the Institutional Review Boards (IRB) at the University of Washington, all referral centers, and at the University of Illinois at Chicago. The University of Florida IRB recognized this de-identified data set as exempt.

Measures

Pain was assessed using the McGill Pain Questionnaire (MPQ).³¹ Participants were asked to rate their pain during the past 24 hours on a scale from 0–10 for pain at its worst, least, and right now. Additionally, participants were also asked to rate their worst common pains including headache, toothache, and stomachache, on a 0–10 scale. In early study waves, pain was measured on a 0–5 descriptor scale, which was subsequently converted to a 0–10 scale based on average pain intensity attributed to each descriptor by a randomly selected sample of hospitalized patients.³²

Participants’ current cancer stage and type of cancer were recorded from medical record review. Types of cancer included breast, colorectal, gastrointestinal, genital-urinary, head and neck, lung, pancreas, prostate, and other primary cancers. For the purposes of this analysis, colorectal, gastrointestinal, and pancreatic cancer were categorized as gastrointestinal cancer. Genital-urinary and prostate cancer were categorized as genital-urinary cancer.

Sociodemographic factors including age, race/ethnicity, gender, education, and study location were collected. Race/ethnicity was categorized into three groups: non-Hispanic White, non-Hispanic Black, and Hispanic. Participants who reported being Asian, Hawaiian Pacific Islander, or others (n=74) were not included in this analysis. Education was dichotomized into three groups: greater than high school to high school diploma, some college to greater than some college, and missing due to non-response. Study location/enrollment year was dichotomized based on the location of participant enrollment: Seattle where data were collected 1994–2002 and Chicago where data were collected 2003–2018.

Data Analysis

First, we described sociodemographic characteristics for all participants. Next, we utilized ANOVAs to assess the relationships between sociodemographic and clinical characteristics associated with worst pain, least pain, and current pain intensity. We then used a generalized linear regression to distinguish the relationship between predictors and worst pain intensity. Statistically significant predictors (p<0.05) at the bivariate level were included in the final model. Study location was included in the final model to adjust for potential confounding attributed to year of participant enrollment. This analysis was limited to variables collected during the baseline visit. All data analyses were conducted using SAS™ (Statistical Analysis Software) version 9.4 (SAS Institute Inc., Cary, NC).

Results

Demographic Factors

Our final sample consisted of 1,339 participants. Table 1 shows a summary of the sociodemographic, cancer stage, type of cancer, and other characteristics of study participants. The average age of participants was 59.4 (SD=14.4) years. Our sample was majority White (61.6%), male (54.7%), and attended some college or greater (48.9%). Regarding cancer factors, 70.8% of participants were diagnosed with stage 4 cancer, 27.7% were diagnosed with lung cancer, 17.6% genital-urinary cancer, 16.9% gastrointestinal cancer, 14.7% head & neck cancer, 11.7 % breast cancer, and 11.3% other types of cancer.

Table 1.

Baseline Demographic Characteristics of Study Population (N=1280)

Characteristic	N (%)
Age- Mean (SD)	59.4 (14.4)
Sex
Female	581 (45.4)
Male	699 (54.6)
Race/Ethnicity
Non-Hispanic White	816 (64.5)
Non-Hispanic Black	358 (28.3)
Hispanic	91 (7.2)
Education
High School or Less	592 (46.3)
Some College – Doctoral Degree	615 (48.1)
Missing	73 (5.7)
Location/Enrollment
Seattle (1994–2002)	598 (46.7)
Chicago (2003–2018)	682 (52.3)
Cancer Stage
0–3	354 (29.5)
4	847 (70.5)
Cancer Site
GI	216 (17.2)
GU	225 (17.9)
Breast	144 (11.5)
Head & Neck	182 (14.5)
Lung	349 (27.8)
Other	142 (11.3)
Common Pain – Mean (SD)
Toothache	7.51 (3.01)
Headache	6.52 (3.12)
Stomachache	6.57 (2.98)
Cancer Pain – Mean (SD)
Worst Pain Intensity	6.60 (2.50)
Least Pain Intensity	2.65 (2.40)
Pain Now	3.74 (2.80)

Bivariate Associations of Worst Pain Intensity

Associations between all study characteristics and worst cancer pain are reported in Table 2. Hispanic participants reported higher mean worst pain (7.23 SD=2.69) than non-Hispanic Black (7.04 SD=2.69) and non-Hispanic White (6.29 SD=2.37) participants. Participants with a high school diploma or less reported higher mean worst pain (6.90 SD=2.51) compared to participants with at least some college (6.36 SD=2.42) and participants who declined to respond (6.20 SD=2.67). Those enrolled in Chicago reported significantly higher mean worst pain (6.95 SD=2.60) than Seattle participants (6.21 SC=2.31). The mean worst cancer pain was higher for those with advanced cancer stage (6.76 SD=2.50) compared to early cancer stage (6.22 SD=2.43). Participants with lung cancer reported greater mean worst pain (6.93SD=2.49) than GI cancers (6.84 SD=2.58), other types of cancers (6.65 SD=2.47), breast cancer (6.62 SD=2.31), head & neck cancer (6.20 SD=2.39), and GU cancer (6.17 SD=2.54).

Table 2.

Bivariate Differences of Worst Least and Current Pain Among Participants (N=1280)

Characteristic	Pain Worst Mean (SD)	P Value		Pain Least Mean (SD)	P-Value		Pain Now Mean (SD)	P-Value
Sex		0.08			<0.01			<0.01
Female (N= 600)	6.73 (2.48)			2.96 (2.54)			4.12 (2.75)
Male (N=728)	6.49 (2.50)			2.44 (2.30)			3.47 (2.80)
Race/Ethnicity		<0.01			<0.01			<0.01
Non-Hispanic White N=813	6.29 (2.37)			2.17 (2.02)			3.11 (2.48)
Non-Hispanic Black N=351	7.04 (2.69)			3.47 (2.85)			4.70 (3.06)
Hispanic =91	7.23 (2.52)			3.41 (2.55)			4.90 (2.71)
Education		<0.01			<0.01			<0.01
High school or Less (N=605)	6.90 (2.51)			2.71 (2.52)			4.00 (3.00)
Some College – Doctoral Degree (N=648)	6.36 (2.42)			2.53 (2.28)			3.44 (2.54)
Missing (N=75)	6.20 (2.67)			3.63 (2.66)			4.68 (2.73)
Location / Enrollment Year		<0.01			<0.01			<0.01
Seattle / (1994-2002) (N=632)	6.21 (2.31)			1.93 (1.92)			2.76 (2.36)
Chicago / (2003–2018) (N=696)	6.95 (2.60)			3.35 (2.63)			4.66 (2.84)
Cancer Stage		<0.01			0.02			<0.01
0–3 (N=366)	6.22 (2.43)			2.37 (2.26)			3.21 (2.71)
4 (N=878)	6.76 (2.50)			2.73 (2.44)			3.93 (2.79)
Type of Cancer		<0.01			<0.01			<0.01
GI (N=217)	6.84 (2.58)			3.48 (2.65)			4.81 (2.84)
GU (N=231)	6.17 (2.54)			2.20 (2.20)			3.20 (2.71)
Breast (N=154)	6.62 (2.31)			3.45 (2.43)			4.36 (2.58)
Head & Neck (N=192)	6.20 (2.39)			2.31 (2.10)			3.03 (2.47)
Lung (N=364)	6.93 (2.49)			2.09 (2.11)			3.30 (2.84)
Other (N=147)	6.65 (2.47)			3.21 (2.63)			4.49 (2.64)

Multivariate Associations of Worst Pain Intensity

In multivariate analyses, race/ethnicity, toothache pain, headache pain, and stomachache pain were statistically significant predictors of worst pain (Table 3). Specifically, being Hispanic (β=0.6759) was associated with higher worst pain intensity than being non-Hispanic White. Having GU cancer (β=−0.6074) was negatively associated with worst cancer pain compared to cancers at other sites. Prior toothache pain (β=0.1191), headache pain (β=0.0549), and stomachache pain (β=0.0578) were also associated with worst pain intensity.

Table 3:

Generalized linear regression coefficients predicting worst pain intensity (N=1280)

Characteristics	Estimate	P-Value
Sex
Female	−0.040	0.82
Male	0.0000	·
Race/Ethnicity
Black	0.3393	0.0980
Hispanic	0. 6759	0.0298
White	0.0000	·
Education
High school or less	0.2917	0. 0674
Missing	−0.5784	0.1524
Some College	0.0000	·
Location
Chicago	0.2980	0.1705
Seattle	0.0000	·
Cancer Stage
4	0.1862	0.3108
0 – 3	0.0000	·
Cancer Site
GI	−0.1935	0.5483
GU	−0. 6074	0.0503
Breast	−0.0909	0.7995
Head & Neck	−0.3043	0.3666
Lung	0. 1557	0.6028
Other	0.0000	·
Characteristics	Estimate	P-Value
Common Pain
Toothache	0. 0961	0.0006
Headache	0. 0549	0. 0423
Stomachache	0.0577	0. 0407

Discussion

In this cross-sectional study, we examined sociodemographic and clinical characteristics associated with worst pain intensity among cancer patients. After controlling for selected covariates, being Hispanic was statistically associated with worst pain intensity compared to non-Hispanic White ethnicity and being non-Hispanic Black approached significance. On the 010 scale, the differences of more than 1 point in worst pain were clinically important. This is one of the few studies identifying Hispanic identity as being positively correlated with cancer pain while controlling for additional sociodemographic and clinical correlates of cancer pain. As researchers continue to investigate racial/ethnic disparities in pain, future studies should continue to investigate disparities within the Hispanic and non-Hispanic Black communities. Additionally, after controlling for selected covariates, our analysis identified previous common pains as statistically significant predictors of worst pain intensity. To our knowledge, this is one of the few studies to identify prior toothache pain intensity as a significant predictor of pain severity among cancer patients. Thus, we identified an underreported marker of worst pain intensity. Furthermore, after controlling for selected covariates our analysis did not find an association between year of enrollment and worst pain intensity, which is surprising given the time span.

Cancer related pain is a highly subjective sensation influenced by several factors including tumor location, stage, and treatment regimen.³³ Historically, clinicians have prescribed opioids or encouraged the use of non-steroidal anti-inflammatory drugs (NSAIDs) to manage pain among cancer patients.^34–37 Previous studies noted opioids were effective in managing mild to severe pain among cancer patients able to tolerate the side effects. However, roughly 32% of cancer patients reported not receiving pharmaceutical pain medications relative to their pain severity.³⁸ Furthermore, approximately only 48% of patient received adequate pharmaceutical pain medications before death.³⁹ Therefore, earlier interventions may improve pain related outcomes among cancer patients.

Sociodemographic Pain Characteristics

Notably, Hispanic participants reported statistically significantly greater pain intensities than non-Hispanic White participants. There are several potential explanations for this finding. There may be potential cultural and language barriers that may affect how pain is communicated between Hispanic patients and providers or researchers. ^40,41 Furthermore, Hispanic individuals are more likely to be screened for cancers at more advanced stages than non-Hispanic White individuals.^42,43 This may result in Hispanic individuals experiencing higher pain intensity than other racial groups. As such, care providers and researchers should continue to work toward reducing barriers to health promotion and cancer screening within Hispanic communities. Further, there are several reasons Hispanic individuals may report greater pain intensity compared to other ethnic groups. Juarez et al., note Hispanic men may be reluctant to report pain as a result of machismo, which may encourage men minimize painful stimuli. Additionally, Hispanic participants may be more likely to understate their pain intensity due to prioritizing family over their own cancer pain.⁴⁴ Similar issues may apply to non-Hispanic Black individuals, although worst pain intensity among non-Hispanic Black participants was not statistically significant in our study.^45,46

Current guidelines suggest using opioids to treat cancer pain among Hispanic populations. However, providers may be reluctant to prescribe opioids to minority patients due to stereotypes, which may have resulted in lower prescription rates among minority groups.^47,48 In addition, providers may inadequately assess pain due to limited knowledge of pain tolerance or a lack of formal training in pain management. For example, a similar study noted minority cancer patients were more likely to report not feeling comfortable discussing their pain with doctors.^49,50 This finding reflects the need for providers to better understand their minority patients’ attitudes, beliefs, and perceptions of pain and to consider use of mobile devices for patients to report their pain.⁵¹

Common Pains and Worst Pain Intensity

Our analysis identified prior worse toothache pain as a significant predictor of worst pain intensity in our population. One possible explanation for this finding may be attributed to the association between toothache and presenting symptoms of renal and breast cancer. Akdas et al., noted several renal cancer patients having significant upper jaw pain and bleeding when diagnosed with renal cell carcinoma.⁵² A similar study noted older cancer patients presenting with secondary breast malignancy also presented with significant toothache and stabbing pain.⁵³ Berry et al., also identified worst toothache pain as a significant correlate of cancer pain after controlling for worst stomachache and headache pain.²⁷ Therefore, to reduce the overall pain intensity reported by individuals living with cancer, clinicians should continue to monitor the oral health of their patients.⁵⁴

Time and Pain Intensity

After controlling for selected covariates, our study noted year of enrollment was not a significant correlate of worst pain intensity. Enrollment for our study spanned 25 years with earlier participants enrolled from Seattle and later participants enrolled from Chicago. After controlling for the year or enrollment, operationalized as study location, we did not note statistical differences between the participants’ enrollment year and worst pain. There are several potential explanations for this finding. First, pain is a highly subjective multifactorial experience that may be difficult to treat due to related biopsychosocial factors.⁵⁵ Further research is needed to identify potential synergistic effects of biological, psychological, and sociological factors including age, mental health, and catastrophizing within cancer patients. Additionally, our finding may be reflective of persistent racial disparities in pain.⁵⁶ For example, a similar study noted patients reporting Hispanic heritage reported higher barrier scores than non-Hispanic patients.⁵⁷ Although clinicians and researchers have made significant advancements in pain treatment, implementation research is needed to target underrepresented and underserved groups most vulnerable to severe pain.

Limitations

Our analysis contained several strengths including a large sample size, validated pain measures, and a significant number of participants from underrepresented ethnic groups. However, there are a few limitations to consider. Initially, this study utilized a cross-sectional design; therefore, we cannot determine the temporality between pain and the predictor variables. Secondly, we only used measures of pain targeting pain intensity and did not include broader aspects of pain including pain interference. Therefore, future studies targeting racial/ethnic disparities should incorporate measures of pain interference as well as pain treatment measures.

Conclusion

As cancer survival rates continue to improve, chronic cancer-associated pain may become an increasing public health concern. Our analysis identified Hispanic identity, prior toothache, stomachache, and headache pain as significant correlates of worst pain intensity. After adjusting for selected covariates, time was not statistically associated with worst pain intensity. Thus clinicians, researchers, and policymakers should continue investigating strategies to reduce disparities among underserved communities and to reduce the burden of worst cancer-related pain and common pains among cancer patients.